Security context

What an agent needs to avoid regressing past fixes and find the next vuln in this repo.

apache/tika

              main @ 29cdf6a
            

Fixes

CVEs

CRITICAL

Peak severity

50.0%

Coverage

Highlights

Denial of Service: 1 prior fix. Scrutinize any change in this area.

tika-parent: most-fixed (1 issue). Treat as high-risk during review.

1 high-severity fix in this history; regressions here are high-impact.

Recurring patterns

The bug types that recur here, drawn from past fixes, not open vulnerabilities.

XML External Entity (XXE) Injection: Tika parsers repeatedly suffer from XXE vulnerabilities (such as CVE-2025-66516, CVE-2025-54988, and CVE-2021-33813) when handling nested or embedded XML documents like PDF XFA or OOXML. This is often driven by improper configuration or reuse of SAXParser instances which inadvertently disables entity expansion limits.

Denial of Service: Unbounded processing of custom metadata properties can cause CPU and memory exhaustion. In SAXBasedMetadataExtractor, passing unlimited text lengths representing decimal properties to BigDecimal allows attackers to trigger a Denial of Service.

HTTP Request Smuggling: Using outdated web server dependencies like Jetty in parent configurations exposes the server component to HTTP/1.1 request smuggling. Upgrading core web dependencies is critical to prevent front-end/back-end desynchronization.